50 
MONOCOTYLEDONES. 
Cyperaceae. — Stems solid and angular ; fruit crustaceous 
or bony. 
Graminaceae. — Steins fistular ; fruit-coat membranous, 
scarce distinguishable from the seed. 
The circle is completed by returning to the spadiceous 
division, of which the alliance to graminaceous plants is 
manifested by the approximation of calamus amongst the 
palms to bambusa of the latter. 
Notwithstanding its unavoidable imperfections, I believe 
the foregoing arrangement will nearly enable any person who 
has no previous knowledge of botany, except understanding 
the terms used, to discover readily the natural order to which 
any monocotyledonous plant before him may belong. If 
my knowledge of all the individuals that compose those or- 
ders had been perfect, I should have distinguished each of 
them from others in the same subsection by one sure feature, 
the most obvious and easy to ascertain that could be found ; 
and I should have subjoined in italics, as auxiliar, all other 
features that seemed to be invariable. But I have to deal 
with orders, some of which are inaccurately constituted, and 
require to be reformed, and of which important features have 
been either overlooked or but partially ascertained ; and I 
have to wade through a long and unsatisfactory string of al- 
ternatives, from which no certain and invariable distinction 
can be elicited. I feel no hesitation in saying that all such 
characters require to be remoulded. Alternatives may 
severally head the sections of an order, but cannot together 
distinguish one order from another. For instance, the words 
“ tube short or long , limb flat or erect,” furnish points for 
subdivision, but no point of distinction for the whole, ex- 
cept the existence of a tube besides the limb, which would be 
properly expressed by the words perianth tubed. In the cha- 
racter of the order from which those words are taken, I find, 
also, embryo orthotropous, heterotropous , or antitropous, from 
which it results, either that the order embraces things so 
dissimilar as to require reformation, or that the position of 
the embryo is immaterial ; but such alternatives, presenting 
every possible position of the embryo, can furnish no peculiar 
distinction of the thing described, and are worse than super- 
fluous, for they so encumber the real marks of separation, 
that it becomes almost impossible to discover them. 
With respect to the suborders I suggest for Liliaceae, it 
must be observed, that it is immaterial whether Melanthaceae 
